24,800 reputation
14992
bio website thehungersite.com
location United Kingdom
age 46
visits member for 3 years, 11 months
seen 11 hours ago
experienced software engineer with many years in the industry, mostly c++ for large-scale, high-reliability systems.

Feb
26
answered Reducing the complexity of over-designed code
Feb
26
comment Is the potential performance hit from LINQ worth it for improved readability?
@HLGEM as I found recently, EF happily added lots of objects, but when it was executed as SQL.. it turned into a lot of SQL inserts. Changing it to use SQLBulkInsert was significantly faster. Devs tend also to just return the entire object rather than selecting the bits they need - that turns into a large query, and lots of unused data transfer. Easy to use often means 'unoptimised'.
Feb
26
answered Authenticating users across multiple web services
Feb
26
answered Observer pattern: Prevent multiple firings on changed internal property
Feb
25
comment Tracing an object
I only mentioned it as Ultimate is very, very.. very expensive.
Feb
25
answered Feasibility of idea to reduce email spam
Feb
25
answered proper way to design REST interface with nginx lua.
Feb
25
revised Comparing path strings against list of rules
added 661 characters in body
Feb
25
comment Tracing an object
Doxygen will map your dependancies, callers and callees and unlike VS Ultimate, its free.
Feb
25
answered Comparing path strings against list of rules
Feb
25
comment Concurrent collections, should read methods allow multiple threads at one time?
@AK_ who said to use this as the lock? I suggested using a CriticalSection or Mutex or similar rather than a private object as the OP is using. I think that's a hangover Java construct.
Feb
25
answered Concurrent collections, should read methods allow multiple threads at one time?
Feb
25
answered Write a (good) code convention for c#
Feb
25
awarded  Good Answer
Feb
24
answered Best way to offer team member help?
Feb
24
comment Is it premature optimization to add database indices?
@SteveJessop cool. I don't think the requirements for DB tables change much without it becoming a different table in all but name. Once you've found the right ID you should keep using it throughout the app, not querying on surname because that's what you have to hand. You should get the ID and query on that instead (for example) but too often people simply add a new index because they didn't write the query to suit the table, they wrote it to suit themselves at the time ... and you end up with indexes on everything. At least that's my take on the question.
Feb
24
comment Is it premature optimization to add database indices?
@SteveJessop I mean you index according to the primary columns that you want to be accessing. For a person table, you might have a search function (if you forget your username you might search on email for example) but afterwards you always use the ID. So ID is the only one that needs indexing. If you do a lot of searching on other fields you might want an index, this will come out in time, but generally you don't want to index every column just because someone sometime decided to write a non-standard query, but you might utilise a different mechanism for these "one-off" cases.
Feb
24
comment Is it premature optimization to add database indices?
@DocBrown I'm not so sure, when you design a table you have (or should have) some understanding how its going to be used. A person table will be queried by ID, or possibly surname. If someone starts to access via DoB, address or phone number then you're going to add indexes for every field - and where does that end?!
Feb
24
awarded  Nice Answer
Feb
24
answered Is it premature optimization to add database indices?